diff --git a/ChangeLog.md b/ChangeLog.md --- a/ChangeLog.md +++ b/ChangeLog.md @@ -5,9 +5,16 @@ ## Unreleased changes +## Version 0.7.0.1 + +- Random 1.2: Parallel functions now require an \`IOGenM\` random number + generator. +- Fix splitting of the random number generator. + + ## Version 0.6.1.1 -- Remove plotting functionality (gnuplot incompatible with ghc921). +- Remove plotting functionality (gnuplot incompatible with ghc922). - Read files strictly. - Refactor; flatten model hierarchy. diff --git a/README.md b/README.md --- a/README.md +++ b/README.md @@ -2,7 +2,7 @@ # The ELynx Suite -Version: 0.6.1.1. +Version: 0.7.0.0. Reproducible evolution made easy.

@@ -74,9 +74,9 @@ # OR: slynx --help Up to date - ELynx Suite version 0.6.1.1. + ELynx Suite version 0.7.0.0. Developed by Dominik Schrempf. - Compiled on February 22, 2022, at 15:10 pm, UTC. + Compiled on July 10, 2022, at 08:48 am, UTC. Usage: slynx [-v|--verbosity VALUE] [-o|--output-file-basename NAME] [-f|--force] [--no-elynx-file] COMMAND @@ -144,16 +144,16 @@ # OR: slynx simulate --help Up to date - ELynx Suite version 0.6.1.1. + ELynx Suite version 0.7.0.0. Developed by Dominik Schrempf. - Compiled on February 22, 2022, at 15:10 pm, UTC. + Compiled on July 10, 2022, at 08:48 am, UTC. Usage: slynx simulate (-t|--tree-file Name) [-s|--substitution-model MODEL] [-m|--mixture-model MODEL] [-e|--edm-file NAME] [-p|--siteprofile-files NAMES] [-w|--mixture-model-weights "[DOUBLE,DOUBLE,...]"] [-g|--gamma-rate-heterogeneity "(NCAT,SHAPE)"] - (-l|--length NUMBER) [-S|--seed [INT]] + (-l|--length NUMBER) [-S|--seed INT] Simulate multi sequence alignments. Available options: @@ -175,8 +175,7 @@ -g,--gamma-rate-heterogeneity "(NCAT,SHAPE)" Number of gamma rate categories and shape parameter -l,--length NUMBER Set alignment length to NUMBER - -S,--seed [INT] Seed for random number generator; list of 32 bit - integers with up to 256 elements (default: random) + -S,--seed INT Seed for random number generator (default: random) -h,--help Show this help text Substitution models: diff --git a/app/Main.hs b/app/Main.hs --- a/app/Main.hs +++ b/app/Main.hs @@ -1,7 +1,7 @@ -- | -- Module : Main -- Description : Work with molecular sequence data --- Copyright : (c) Dominik Schrempf 2021 +-- Copyright : 2021 Dominik Schrempf -- License : GPL-3.0-or-later -- -- Maintainer : dominik.schrempf@gmail.com diff --git a/slynx.cabal b/slynx.cabal --- a/slynx.cabal +++ b/slynx.cabal @@ -1,54 +1,54 @@ -cabal-version: 3.0 -name: slynx -version: 0.6.1.1 -synopsis: Handle molecular sequences -description: Examine, modify, and simulate molecular sequences in a reproducible way. Please see the README on GitHub at . -category: Bioinformatics -homepage: https://github.com/dschrempf/elynx#readme -bug-reports: https://github.com/dschrempf/elynx/issues -author: Dominik Schrempf -maintainer: dominik.schrempf@gmail.com -copyright: Dominik Schrempf (2021) -license: GPL-3.0-or-later -license-file: LICENSE -build-type: Simple +cabal-version: 3.0 +name: slynx +version: 0.7.0.1 +synopsis: Handle molecular sequences +description: + Examine, modify, and simulate molecular sequences in a reproducible way. Please see the README on GitHub at . +category: Bioinformatics +homepage: https://github.com/dschrempf/elynx#readme +bug-reports: https://github.com/dschrempf/elynx/issues +author: Dominik Schrempf +maintainer: dominik.schrempf@gmail.com +copyright: 2021 Dominik Schrempf +license: GPL-3.0-or-later +license-file: LICENSE +build-type: Simple extra-source-files: - README.md - ChangeLog.md + ChangeLog.md + README.md source-repository head - type: git + type: git location: https://github.com/dschrempf/elynx library exposed-modules: - SLynx.Concatenate.Concatenate - SLynx.Concatenate.Options - SLynx.Examine.Examine - SLynx.Examine.Options - SLynx.Filter.Filter - SLynx.Filter.Options - SLynx.Options - SLynx.Simulate.Options - SLynx.Simulate.PhyloModel - SLynx.Simulate.Simulate - SLynx.SLynx - SLynx.SubSample.Options - SLynx.SubSample.SubSample - SLynx.Tools - SLynx.Translate.Options - SLynx.Translate.Translate - other-modules: - Paths_slynx - autogen-modules: - Paths_slynx - hs-source-dirs: src - ghc-options: -Wall -Wunused-packages + SLynx.Concatenate.Concatenate + SLynx.Concatenate.Options + SLynx.Examine.Examine + SLynx.Examine.Options + SLynx.Filter.Filter + SLynx.Filter.Options + SLynx.Options + SLynx.Simulate.Options + SLynx.Simulate.PhyloModel + SLynx.Simulate.Simulate + SLynx.SLynx + SLynx.SubSample.Options + SLynx.SubSample.SubSample + SLynx.Tools + SLynx.Translate.Options + SLynx.Translate.Translate + + other-modules: Paths_slynx + autogen-modules: Paths_slynx + hs-source-dirs: src + ghc-options: -Wall -Wunused-packages build-depends: - aeson + , aeson , attoparsec - , base >=4.7 && <5 + , base >=4.7 && <5 , bytestring , containers , elynx-markov @@ -56,23 +56,26 @@ , elynx-tools , elynx-tree , hmatrix - , mwc-random + , matrices , optparse-applicative + , random , statistics , text , transformers , vector + default-language: Haskell2010 executable slynx - main-is: Main.hs - other-modules: - Paths_slynx - autogen-modules: - Paths_slynx - hs-source-dirs: app - ghc-options: -Wall -Wunused-packages -threaded -rtsopts -with-rtsopts=-N + main-is: Main.hs + other-modules: Paths_slynx + autogen-modules: Paths_slynx + hs-source-dirs: app + ghc-options: + -Wall -Wunused-packages -threaded -rtsopts -with-rtsopts=-N + build-depends: - base >=4.7 && <5 + , base >=4.7 && <5 , slynx + default-language: Haskell2010 diff --git a/src/SLynx/Concatenate/Concatenate.hs b/src/SLynx/Concatenate/Concatenate.hs --- a/src/SLynx/Concatenate/Concatenate.hs +++ b/src/SLynx/Concatenate/Concatenate.hs @@ -4,7 +4,7 @@ -- | -- Module : Analyze.Analyze -- Description : Parse sequence file formats and analyze them --- Copyright : (c) Dominik Schrempf 2021 +-- Copyright : 2021 Dominik Schrempf -- License : GPL-3.0-or-later -- -- Maintainer : dominik.schrempf@gmail.com diff --git a/src/SLynx/Concatenate/Options.hs b/src/SLynx/Concatenate/Options.hs --- a/src/SLynx/Concatenate/Options.hs +++ b/src/SLynx/Concatenate/Options.hs @@ -3,7 +3,7 @@ -- | -- Module : SLynx.Concatenate.Options -- Description : ELynxSeq argument parsing --- Copyright : (c) Dominik Schrempf 2021 +-- Copyright : 2021 Dominik Schrempf -- License : GPL-3.0-or-later -- -- Maintainer : dominik.schrempf@gmail.com diff --git a/src/SLynx/Examine/Examine.hs b/src/SLynx/Examine/Examine.hs --- a/src/SLynx/Examine/Examine.hs +++ b/src/SLynx/Examine/Examine.hs @@ -5,7 +5,7 @@ -- -- Module : Analyze.Analyze -- Description : Parse sequence file formats and analyze them --- Copyright : (c) Dominik Schrempf 2021 +-- Copyright : 2021 Dominik Schrempf -- License : GPL-3.0-or-later -- -- Maintainer : dominik.schrempf@gmail.com @@ -18,21 +18,28 @@ ) where +import Control.Monad +import Control.Monad.IO.Class import Control.Monad.Trans.Reader import qualified Data.ByteString.Lazy.Char8 as BL +import qualified Data.Matrix.Unboxed as MU import qualified Data.Set as S import qualified Data.Vector.Unboxed as V import qualified ELynx.Alphabet.Alphabet as A import qualified ELynx.Alphabet.Character as C import qualified ELynx.Sequence.Alignment as M import qualified ELynx.Sequence.Distance as D +import ELynx.Sequence.Divergence import qualified ELynx.Sequence.Sequence as Seq import ELynx.Tools.ByteString import ELynx.Tools.ELynx import ELynx.Tools.Environment +import ELynx.Tools.Logger +import qualified Numeric.LinearAlgebra as L import SLynx.Examine.Options import SLynx.Tools import qualified Statistics.Sample as Sm +import System.IO import Text.Printf pRow :: String -> String -> BL.ByteString @@ -44,30 +51,29 @@ examineAlignment :: Bool -> M.Alignment -> BL.ByteString examineAlignment perSiteFlag a = BL.unlines - [ BL.pack - "Sequences have equal length (multi sequence alignment, or single sequence).", - BL.pack "Number of columns in alignment:", + [ "Sequences have equal length (multi sequence alignment, or single sequence).", + "Number of columns in alignment:", pRow " Total:" $ show aL, pRow " Constant:" $ show nConstant, pRow " Constant (including gaps or unknowns):" $ show nConstantSoft, pRow " Without gaps:" $ show (M.length aNoGaps), pRow " With standard characters only:" $ show (M.length aOnlyStd), - BL.empty, + "", pRow "Total number of characters:" $ show nTot, pRow "Standard (i.e., not extended IUPAC) characters:" $ show (nTot - nIUPAC - nGaps - nUnknowns), pRow "Extended IUPAC characters:" $ show nIUPAC, pRow "Gaps:" $ show nGaps, pRow "Unknowns:" $ show nUnknowns, - BL.empty, + "", pRow "Percentage of standard characters:" $ printf "%2.2f" (100.0 - percentIUPAC - percentGaps - percentUnknowns), pRow "Percentage of extended IUPAC characters:" $ printf "%2.2f" percentIUPAC, pRow "Percentage of gaps:" $ printf "%2.2f" percentGaps, pRow "Percentage of unknowns:" $ printf "%2.2f" percentUnknowns, - BL.empty, + "", BL.pack "Distribution of characters:", -- Holy crap. BL.pack $ @@ -77,24 +83,26 @@ A.alphabetSpec $ M.alphabet a, BL.pack $ unwords $ map (printf "%.3f") charFreqs, - BL.empty, + "", BL.pack "Pairwise hamming distances (per site):", pRow " Mean:" $ printf "%.3f" hMean, pRow " Standard deviation:" $ printf "%.3f" $ sqrt hVar, pRow " Minimum:" $ printf "%.3f" hMin, pRow " Maximum:" $ printf "%.3f" hMax, - BL.empty, - BL.pack "Mean effective number of states (measured using entropy):", + "", + "Mean effective number of states (measured using entropy):", pRow "Across whole alignment:" $ printf "%.3f" kEffMean, pRow "Across columns without gaps:" $ printf "%.3f" kEffMeanNoGaps, pRow "Across columns without extended IUPAC characters:" $ printf "%.3f" kEffMeanOnlyStd, - BL.empty, + "", BL.pack "Mean effective number of states (measured using homoplasy):", pRow "Across whole alignment:" $ printf "%.3f" kEffMeanHomo, pRow "Across columns without gaps:" $ printf "%.3f" kEffMeanNoGapsHomo, pRow "Across columns without extended IUPAC characters:" $ - printf "%.3f" kEffMeanOnlyStdHomo + printf "%.3f" kEffMeanOnlyStdHomo, + "", + "Divergence matrix:" ] <> perSiteBS where @@ -155,10 +163,34 @@ Left _ -> BL.empty Right a -> BL.pack "\n" <> examineAlignment perSiteFlag a +-- From https://stackoverflow.com/a/52602906/3536806. +tuples :: [a] -> [(a, a)] +tuples [] = [] +tuples (x : xs) = map (\y -> (x, y)) xs ++ tuples xs + +-- This is all ugly, but who cares. +writeDivergenceMatrix :: Handle -> (Seq.Sequence, Seq.Sequence, MU.Matrix Int) -> IO () +writeDivergenceMatrix h (x, y, xs) = do + BL.hPutStrLn h $ "> " <> Seq.name x <> ", " <> Seq.name y + hPutStr h $ L.dispf 0 m + where + n = MU.rows xs + m = L.matrix n $ map fromIntegral $ MU.toList xs + +computeDivergenceMatrices :: [Seq.Sequence] -> ELynx ExamineArguments () +computeDivergenceMatrices ss = do + logInfoS "Compute divergence matrices." + let xys = tuples ss + ds = map (\(x, y) -> (x, y, either error id $ divergence x y)) xys + h <- outHandle "divergence matrices" ".div" + sequence_ $ map (liftIO . writeDivergenceMatrix h) ds + liftIO $ hClose h + -- | Examine sequences. examineCmd :: ELynx ExamineArguments () examineCmd = do - (ExamineArguments al inFile perSiteFlag) <- localArguments <$> ask + (ExamineArguments al inFile perSiteFlag divergenceFlag) <- localArguments <$> ask ss <- readSeqs al inFile let result = examine perSiteFlag ss out "result of examination" result ".out" + when divergenceFlag (computeDivergenceMatrices ss) diff --git a/src/SLynx/Examine/Options.hs b/src/SLynx/Examine/Options.hs --- a/src/SLynx/Examine/Options.hs +++ b/src/SLynx/Examine/Options.hs @@ -3,7 +3,7 @@ -- | -- Module : SLynx.Examine.Options -- Description : ELynxSeq argument parsing --- Copyright : (c) Dominik Schrempf 2021 +-- Copyright : 2021 Dominik Schrempf -- License : GPL-3.0-or-later -- -- Maintainer : dominik.schrempf@gmail.com @@ -28,13 +28,14 @@ data ExamineArguments = ExamineArguments { exAlphabet :: Alphabet, exInFile :: FilePath, - exPerSite :: Bool + exPerSite :: Bool, + exDivergence :: Bool } deriving (Eq, Show, Generic) instance Reproducible ExamineArguments where inFiles = pure . exInFile - outSuffixes _ = [".out"] + outSuffixes args = if exDivergence args then [".out", ".div"] else [".out"] getSeed _ = Nothing -- XXX: Probably throw error when seed is set. @@ -52,11 +53,15 @@ -- | Command line parser. examineArguments :: Parser ExamineArguments examineArguments = - ExamineArguments <$> alphabetOpt <*> filePathArg <*> examinePerSiteOpt + ExamineArguments <$> alphabetOpt <*> filePathArg <*> examinePerSiteOpt <*> examineDivergence examinePerSiteOpt :: Parser Bool examinePerSiteOpt = switch $ long "per-site" <> help "Report per site summary statistics" + +examineDivergence :: Parser Bool +examineDivergence = + switch $ long "divergence" <> help "Compute pairwise divergence matrices" filePathArg :: Parser FilePath filePathArg = diff --git a/src/SLynx/Filter/Filter.hs b/src/SLynx/Filter/Filter.hs --- a/src/SLynx/Filter/Filter.hs +++ b/src/SLynx/Filter/Filter.hs @@ -4,7 +4,7 @@ -- | -- Module : Analyze.Analyze -- Description : Parse sequence file formats and analyze them --- Copyright : (c) Dominik Schrempf 2021 +-- Copyright : 2021 Dominik Schrempf -- License : GPL-3.0-or-later -- -- Maintainer : dominik.schrempf@gmail.com diff --git a/src/SLynx/Filter/Options.hs b/src/SLynx/Filter/Options.hs --- a/src/SLynx/Filter/Options.hs +++ b/src/SLynx/Filter/Options.hs @@ -3,7 +3,7 @@ -- | -- Module : SLynx.Filter.Options -- Description : ELynxSeq argument parsing --- Copyright : (c) Dominik Schrempf 2021 +-- Copyright : 2021 Dominik Schrempf -- License : GPL-3.0-or-later -- -- Maintainer : dominik.schrempf@gmail.com @@ -85,7 +85,8 @@ filterLongerThanOpt = optional $ option auto $ - long "longer-than" <> metavar "LENGTH" + long "longer-than" + <> metavar "LENGTH" <> help "Only keep sequences longer than LENGTH" @@ -93,7 +94,8 @@ filterShorterThanOpt = optional $ option auto $ - long "shorter-than" <> metavar "LENGTH" + long "shorter-than" + <> metavar "LENGTH" <> help "Only keep sequences shorter than LENGTH" diff --git a/src/SLynx/Options.hs b/src/SLynx/Options.hs --- a/src/SLynx/Options.hs +++ b/src/SLynx/Options.hs @@ -3,7 +3,7 @@ -- | -- Module : SLynx.Options -- Description : SLynx general options --- Copyright : (c) Dominik Schrempf 2021 +-- Copyright : 2021 Dominik Schrempf -- License : GPL-3.0-or-later -- -- Maintainer : dominik.schrempf@gmail.com diff --git a/src/SLynx/SLynx.hs b/src/SLynx/SLynx.hs --- a/src/SLynx/SLynx.hs +++ b/src/SLynx/SLynx.hs @@ -1,7 +1,7 @@ -- | -- Module : SLynx.SLynx -- Description : SLynx module --- Copyright : (c) Dominik Schrempf 2021 +-- Copyright : 2021 Dominik Schrempf -- License : GPL-3.0-or-later -- -- Maintainer : dominik.schrempf@gmail.com diff --git a/src/SLynx/Simulate/Options.hs b/src/SLynx/Simulate/Options.hs --- a/src/SLynx/Simulate/Options.hs +++ b/src/SLynx/Simulate/Options.hs @@ -3,7 +3,7 @@ -- | -- Module : SLynx.Simulate.Options -- Description : ELynxSim argument parsing --- Copyright : (c) Dominik Schrempf 2021 +-- Copyright : 2021 Dominik Schrempf -- License : GPL-3.0-or-later -- -- Maintainer : dominik.schrempf@gmail.com @@ -69,8 +69,8 @@ instance Reproducible SimulateArguments where inFiles a = - argsTreeFile a : - (maybeToList (argsEDMFile a) ++ fromMaybe [] (argsSiteprofilesFiles a)) + argsTreeFile a + : (maybeToList (argsEDMFile a) ++ fromMaybe [] (argsSiteprofilesFiles a)) outSuffixes _ = [".model.gz", ".fasta"] getSeed = Just . argsSeed setSeed a s = a {argsSeed = s} @@ -100,7 +100,9 @@ treeFileOpt :: Parser FilePath treeFileOpt = strOption $ - long "tree-file" <> short 't' <> metavar "Name" + long "tree-file" + <> short 't' + <> metavar "Name" <> help "Read tree from Newick file NAME" @@ -129,7 +131,9 @@ maybeEDMFileOpt = optional $ strOption - ( long "edm-file" <> short 'e' <> metavar "NAME" + ( long "edm-file" + <> short 'e' + <> metavar "NAME" <> help "Empirical distribution model file NAME in Phylobayes format" ) @@ -139,7 +143,9 @@ optional $ words <$> strOption - ( long "siteprofile-files" <> short 'p' <> metavar "NAMES" + ( long "siteprofile-files" + <> short 'p' + <> metavar "NAMES" <> help "File names of site profiles in Phylobayes format" ) @@ -170,7 +176,9 @@ lengthOpt = option auto - ( long "length" <> short 'l' <> metavar "NUMBER" + ( long "length" + <> short 'l' + <> metavar "NUMBER" <> help "Set alignment length to NUMBER" ) diff --git a/src/SLynx/Simulate/PhyloModel.hs b/src/SLynx/Simulate/PhyloModel.hs --- a/src/SLynx/Simulate/PhyloModel.hs +++ b/src/SLynx/Simulate/PhyloModel.hs @@ -3,7 +3,7 @@ -- | -- Module : SLynx.Simulate.PhyloModel -- Description : Parse and interpret the model string --- Copyright : (c) Dominik Schrempf 2021 +-- Copyright : 2021 Dominik Schrempf -- License : GPL-3.0-or-later -- -- Maintainer : dominik.schrempf@gmail.com diff --git a/src/SLynx/Simulate/Simulate.hs b/src/SLynx/Simulate/Simulate.hs --- a/src/SLynx/Simulate/Simulate.hs +++ b/src/SLynx/Simulate/Simulate.hs @@ -4,7 +4,7 @@ -- | -- Module : SLynx.Simulate.Simulate -- Description : Simulate multiple sequence alignments --- Copyright : (c) Dominik Schrempf 2021 +-- Copyright : 2021 Dominik Schrempf -- License : GPL-3.0-or-later -- -- Maintainer : dominik.schrempf@gmail.com @@ -56,7 +56,7 @@ import qualified Numeric.LinearAlgebra as L import SLynx.Simulate.Options import SLynx.Simulate.PhyloModel -import System.Random.MWC +import System.Random.Stateful import Text.Printf -- Display a vector with given precision. @@ -91,11 +91,11 @@ -- Simulate a 'Alignment' for a given phylogenetic model, -- phylogenetic tree, and alignment length. simulateAlignment :: - (HasLength e, HasName a) => + (RandomGen g, HasLength e, HasName a) => MP.PhyloModel -> Tree e a -> Int -> - GenIO -> + IOGenM g -> ELynx SimulateArguments () simulateAlignment pm t' n g = do let t = fromLength . getLength <$> toTreeBranchLabels t' @@ -182,8 +182,8 @@ summarizeSM :: MS.SubstitutionModel -> [BL.ByteString] summarizeSM sm = map BL.pack $ - (show (MS.alphabet sm) ++ " substitution model: " ++ MS.name sm ++ ".") : - ["Parameters: " ++ show (MS.params sm) ++ "." | not (null (MS.params sm))] + (show (MS.alphabet sm) ++ " substitution model: " ++ MS.name sm ++ ".") + : ["Parameters: " ++ show (MS.params sm) ++ "." | not (null (MS.params sm))] ++ case MS.alphabet sm of DNA -> [ "Stationary distribution: " @@ -208,8 +208,8 @@ summarizeMMComponent :: MM.Component -> [BL.ByteString] summarizeMMComponent c = BL.pack "Weight: " - <> (BB.toLazyByteString . BB.doubleDec $ MM.weight c) : - summarizeSM (MM.substModel c) + <> (BB.toLazyByteString . BB.doubleDec $ MM.weight c) + : summarizeSM (MM.substModel c) -- Summarize a mixture model; lines to be printed to screen or log. summarizeMM :: MM.MixtureModel -> [BL.ByteString] @@ -297,7 +297,7 @@ logInfoS "Simulate alignment." let alignmentLength = argsLength l logInfoS $ "Length: " <> show alignmentLength <> "." - gen <- liftIO <$> initialize $ case argsSeed l of + gen <- newIOGenM $ mkStdGen $ case argsSeed l of RandomUnset -> error "simulateCmd: seed not available; please contact maintainer." RandomSet s -> s Fixed s -> s diff --git a/src/SLynx/SubSample/Options.hs b/src/SLynx/SubSample/Options.hs --- a/src/SLynx/SubSample/Options.hs +++ b/src/SLynx/SubSample/Options.hs @@ -3,7 +3,7 @@ -- | -- Module : SLynx.SubSample.Options -- Description : ELynxSeq argument parsing --- Copyright : (c) Dominik Schrempf 2021 +-- Copyright : 2021 Dominik Schrempf -- License : GPL-3.0-or-later -- -- Maintainer : dominik.schrempf@gmail.com @@ -83,7 +83,9 @@ subSampleNSitesOpt :: Parser Int subSampleNSitesOpt = option auto $ - long "number-of-sites" <> short 'n' <> metavar "INT" + long "number-of-sites" + <> short 'n' + <> metavar "INT" <> help "Number of sites randomly drawn with replacement" diff --git a/src/SLynx/SubSample/SubSample.hs b/src/SLynx/SubSample/SubSample.hs --- a/src/SLynx/SubSample/SubSample.hs +++ b/src/SLynx/SubSample/SubSample.hs @@ -4,7 +4,7 @@ -- | -- Module : Analyze.Analyze -- Description : Parse sequence file formats and analyze them --- Copyright : (c) Dominik Schrempf 2021 +-- Copyright : 2021 Dominik Schrempf -- License : GPL-3.0-or-later -- -- Maintainer : dominik.schrempf@gmail.com @@ -28,16 +28,19 @@ import ELynx.Tools.Reproduction import SLynx.SubSample.Options import SLynx.Tools -import System.Random.MWC +import System.Random.Stateful -- | Sub sample sequences. subSampleCmd :: ELynx SubSampleArguments () subSampleCmd = do - (SubSampleArguments al inFile nSites nAlignments (Fixed s)) <- localArguments <$> ask + (SubSampleArguments al inFile nSites nAlignments sOpt) <- localArguments <$> ask + let s = case fromSeedOpt sOpt of + Nothing -> error "subSampleCmd: No seed." + Just x -> x logInfoS $ " Sample " <> show nSites <> " sites." logInfoS $ " Sample " <> show nAlignments <> " multi sequence alignments." ss <- readSeqs al inFile - gen <- liftIO $ initialize s + gen <- newIOGenM $ mkStdGen s let a = either error id (M.fromSequences ss) samples <- liftIO $ replicateM nAlignments $ M.randomSubSample nSites a gen let results = map (sequencesToFasta . M.toSequences) samples diff --git a/src/SLynx/Tools.hs b/src/SLynx/Tools.hs --- a/src/SLynx/Tools.hs +++ b/src/SLynx/Tools.hs @@ -4,7 +4,7 @@ -- | -- Module : SLynx.Tools -- Description : Common tools for sequence lynx --- Copyright : (c) Dominik Schrempf 2021 +-- Copyright : 2021 Dominik Schrempf -- License : GPL-3.0-or-later -- -- Maintainer : dominik.schrempf@gmail.com @@ -48,6 +48,8 @@ alphabetOpt :: Parser Alphabet alphabetOpt = option auto $ - long "alphabet" <> short 'a' <> metavar "NAME" + long "alphabet" + <> short 'a' + <> metavar "NAME" <> help "Specify alphabet type NAME" diff --git a/src/SLynx/Translate/Options.hs b/src/SLynx/Translate/Options.hs --- a/src/SLynx/Translate/Options.hs +++ b/src/SLynx/Translate/Options.hs @@ -3,7 +3,7 @@ -- | -- Module : SLynx.Translate.Options -- Description : ELynxSeq argument parsing --- Copyright : (c) Dominik Schrempf 2021 +-- Copyright : 2021 Dominik Schrempf -- License : GPL-3.0-or-later -- -- Maintainer : dominik.schrempf@gmail.com @@ -60,14 +60,18 @@ readingFrameOpt :: Parser Int readingFrameOpt = option auto $ - long "reading-frame" <> short 'r' <> metavar "INT" + long "reading-frame" + <> short 'r' + <> metavar "INT" <> help "Reading frame [0|1|2]." universalCodeOpt :: Parser UniversalCode universalCodeOpt = option auto $ - long "universal-code" <> short 'u' <> metavar "CODE" + long "universal-code" + <> short 'u' + <> metavar "CODE" <> help ("universal code; one of: " ++ codeStr ++ ".") where diff --git a/src/SLynx/Translate/Translate.hs b/src/SLynx/Translate/Translate.hs --- a/src/SLynx/Translate/Translate.hs +++ b/src/SLynx/Translate/Translate.hs @@ -4,7 +4,7 @@ -- | -- Module : Analyze.Analyze -- Description : Parse sequence file formats and analyze them --- Copyright : (c) Dominik Schrempf 2021 +-- Copyright : 2021 Dominik Schrempf -- License : GPL-3.0-or-later -- -- Maintainer : dominik.schrempf@gmail.com